News: Event News

On November 10th, 2023, the CMU Portugal Program hosted the Talk “Experiments with Social Coordination in Hybrid Systems of Humans and Bots” with Hirokazu Shirado, Assistant Professor of the Human-Computer Interaction Institute (HCII Institute) in the School of Computer Science at Carnegie Mellon University (CMU). This event took place at the CMU Portugal/ITI Premises at Factory (Hub Criativo do Beato) and drew an engaged audience of over 20 participants.

Shirado’s expertise lies in investigating the interplay between social interactions and technology, particularly in understanding how social order emerges and how communities collectively navigate social dilemmas. His thought-provoking presentation shed light on experimental studies exploring the dynamics of human coordination within the context of human-AI hybrid systems, offering valuable insights into the potential impact on human prosociality.

During the talk, Shirado delved into groundbreaking research, including a study published in Nature in 2017. This research involved introducing bots into experimental networks where individuals were engaged in a coordination game, revealing that the performance of human groups improved when bots exhibited seemingly irrational behaviors.

The second part of the presentation focused on a recent study investigating the social implications of intelligent assistance, such as emergency automatic steering assistance, in a novel cyber-physical experiment. The findings prompted a compelling discussion on how AI could potentially influence human prosociality by altering social foci.

Hirokazu Shirado was in Portugal for two months as a CMU Portugal Visiting Faculty and Researcher collaborating with Prof. Ana Paiva at INESC ID I Técnico GAIPS Lab (Research Group on AI for People and the Society”): “My two-month visit to Portugal has been an amazing opportunity for me to expand my research agenda and academic networks. I’ve had the chance to exchange ideas with various researchers at Portuguese institutions, including Ana Paiva and her lab members. Additionally, my stay in Portugal has allowed me to learn more about Portuguese culture, which has led me to reflect on my research interests in human cooperation across multiple cultures”, shares Shirado. 

During his stay in Portugal, he had the chance to discuss with Ana Paiva about the potential of machines to facilitate human collaboration and decision-making in groups. While machine intelligence has been developed for individual use, it may not always be effective in group settings due to social dynamics and network mechanisms. 

“To tackle this challenge, Ana and I have come up with a research idea that involves robot control to manage the attention dynamics in communication networks between humans and robots. This idea is a combination of Ana’s social robot studies and my social network experiments on resource allocation. We are currently discussing the details of the research project and plan to execute it once I return to Pittsburgh”, explains the researcher.”

Learn more about the speaker and his research here.  

 

The CMU Portugal Program hosted a Fall Lunch on campus in the Cohon University Center at CMU, on November 8, 2023. This event brought together Dual Degree Ph.D. and Affiliated Ph.D. students currently studying at CMU as well as visitors from the Visiting Students and Visiting Faculty and Researchers Programs. 

The welcome event was hosted by CMU Portugal Director José Moura and counted on members from seven different departments including Electrical Computer Engineering, Engineering and Public Policy, Computer Science, Human Computer Interaction, Robotics, Language Technologies and Software & Societal Systems, providing an opportunity for all  those connected by the CMU Portugal to network and enjoy a meal together as the holiday season begins. 

The fall semester at CMU will be attended by a group of 41 CMU Portugal members, including 17 Dual Degree Ph.D. students, 4 Affiliated Ph.D. students, 11 Visiting Students and 9 visiting faculty and researchers at CMU this semester.

TAMI, a CMU Portugal Large Scale Collaborative Research Project, has been focused on using Artificial  Intelligence (AI) to help make medical diagnosis clearer and more reliable. The three-year initiative has been able to develop a decision-supporting platform based on AI algorithms for the diagnosis of health complications such as cervical cancer, lung diseases and eye diseases. 

In a final event hosted at UPTEC (Porto) on September 14th, the project’s consortia led by First Solutions, in collaboration with Fraunhofer Portugal, INESC TEC and Administração Regional de Saúde do Norte (ARS Norte), shared the project’s main achievements and outcomes. 

In the culmination of three years of dedicated research and development, the Transparent Artificial Medical Intelligence (TAMI) project, a CMU Portugal Large Scale Collaborative Research Project, is set to improve medical diagnosis through the power of Artificial Intelligence (AI).

Led by First Solutions in collaboration with Fraunhofer Portugal, INESC TEC , and Administração Regional de Saúde do Norte (ARS Norte), the TAMI consortium has strived to create a decision-supporting platform that harnesses AI algorithms for diagnosing complex health conditions, including cervical cancer, lung diseases, and eye diseases.

Tiago Oliveira, project’s PI from First Solutions kicked off the meeting with an overall presentation of TAMI’s objectives and accomplishments. TAMI has been mainly focused on the development of tools rooted in Explainable Artificial Intelligence (XAI). These tools are designed to assist healthcare professionals in identifying “relevant findings” within images across various medical domains, including cervical cytology, colposcopy, glaucoma, and chest x-ray.

TAMI’s ultimate goal is to bring this groundbreaking application to market, making it accessible to the healthcare sector,such as hospitals and clinics. This would enable “consumers access to self-explanatory AI decisions for screening and diagnosis, , promising a transformation in medical care. However, despite remarkable progress, TAMI still faces the hurdle of navigating complex legal requirements and obtaining necessary certifications.

The project presentation was followed by a look into the four medical areas studied under TAMI: Colposcopy presented by Jaime Cardoso from INESC TEC, Cervical Cytology by Luís Rosado, Glaucoma by Filipe Soares from Fraunhofer Portugal AICOS, RX Thorax by Ana Maria Mendonça from INESC TEC. 

Representing the research team from Carnegie Mellon University, Asim Smailagic from the Electrical and Computer Engineering Department, presented the Carnegie Mellon contribution to TAMI’s research outcomes.

“The CMU team contributes to TAMI activities 4 and 5: Next Generation of Explainable AI Methods; and Methods for Evaluating Interpretability. CMU contributed to three accepted journal papers and five accepted conference papers, in collaboration with FhP and INESC TEC. CMU fostered collaboration between participating institutions by hosting visiting scholars from our partner institutions, and both scholarly research periods resulting in published works. From our side, we consider that the TAMI project has been a great success. The unique structure of this CMU Portugal project bridges academia (CMU and INESC), with industry (First and FhP), and the hospital (ARSN). Our multi-faceted collaborations are showing long-term benefits that promise to extend beyond the TAMI project itself.”, explains Asim Smailagic. 

The event wrapped up with a demonstration of the TAMI platform’s capabilities. This platform is intended to support medical professionals by providing AI-driven explanations for specific disease diagnoses, alongside visual elements pinpointing relevant areas within images and textual elements that convey comprehensible insights to human users. The platform has two main features: analyzing images to identify areas that may have relevant information and, at the same time, classifying variables depending on the specialty. The application, thoughtfully developed by First Solutions, is set to usher in a new era of medical decision support. In the near future, as the platform advances towards maturity, it will reshape the landscape of healthcare, bridging the gap between cutting-edge AI technology and medical practice.

The TAMI (NORTE-01-0247-FEDER-045905) project was selected by the ‘Go Portugal – Global Science and Technology Partnerships Portugal’ initiative in the context of the CMU Portugal program. It is co-financed by the European Regional Development Fund (ERDF) through the Operational Program for Competitiveness and Internationalization (COMPETE 2020) and by the Portuguese Foundation for Science and Technology (FCT) under the CMU Portugal Program.

In the Media: Sapo Tek

 

The CMU Portugal Program held a Welcome Back event on September 19th addressed to all students spending this next semester in Pittsburgh. From the 23 students currently attending the Fall semester at Carnegie Mellon, 10 attended the event and stopped by to grab coffee & donuts, while meeting other CMU Portugal colleagues.

On Campus  events are an excellent opportunity to network and foster a supportive community of Ph.D. candidates, providing participants the opportunity to connect and share experiences with their peers. 

The Program currently has 23 active students attending the Fall semester at Carnegie Mellon, and anticipates 22 visiting students and faculty later in the semester. 

This year 6 new Dual Degree students have joined the Program and will soon be able to spend two years at Carnegie Mellon.

The 2023 edition of the Lisbon Machine Learning Summer School (LxMLS 2023) was held in July, bringing together more than 170 students for an intensive week focused on exploring the realms of machine learning (ML) and natural language processing (NLP). Once again,  CMU Portugal Program has supported this event, continuing the strong connection with this annual event, which has now spanned over a decade.

The school encompasses a wide spectrum of machine learning topics, ranging from theoretical foundations to practical applications, all of which play a pivotal role in tackling the challenges presented by language processing problems arising in different application areas. The event is a collaborative effort involving Instituto Superior Técnico (Técnico), Instituto de Telecomunicações, Instituto de Engenharia de Sistemas e Computadores, Investigação e Desenvolvimento em Lisboa (INESC-ID), the Lisbon ELLIS Unit for Learning and Intelligent Systems (LUMLIS), and Unbabel, Zendesk, and IBM Research, with the support of the CMU Portugal Program.

The 13th edition drew in a diverse group of participants, including experts in the field of ML and NLP, reaching out to students of various nationalities and academic backgrounds, such as Electrical Engineering, Mathematics, Biomedical Engineering, Linguistics, and Economics. 

Among the speakers at the event was Bhiksha Raj, Carnegie Mellon University Professor at  Language Technologies Institute and  Fellow at IEEE. Bhiksha Raj works in four broad areas: Speech recognition, Audio processing, Neural networks, and Privacy/Security for voice processing.  His involvement in multiple LxMLS editions underscores his commitment to the program. Additionally, he has collaborated with Portuguese researchers on projects backed by the CMU Portugal initiative and has co-supervised dual-degree Ph.D. candidates.

The event also featured the active participation of CMU Portugal faculty and alums in the organization and speaking engagements. They were André Martins, Isabel Trancoso, Mário Figueiredo, and Zita Marinho, contributing to the event’s success. The steering committee also included Fernando Pereira, a member of CMU Portugal’s External Review Committee, who serves as a VP and Engineering Fellow at Google, where hehe leads pioneering research and development efforts in the realms of natural language understanding and machine learning.

On July 14th, the CMU Portugal Large Scale Collaborative Project “Building Hope” concluded its journey with a closing session. This online event brought together more than 50 participants to discuss the project’s latest advancements in energy management.

Funded through PT2020 and FCT under the CMU Portugal Program, Building HOPE – Building Holistic Optimization of Prosumed Energy – aimed to revolutionize energy management practices in smart urban environments. The project focused on the development and validation of an innovative tool known as Holistic Optimization of Prosumed Energy (HOPE) specifically designed for buildings. By leveraging cutting-edge technologies such as data collection, sensing advancements, and AI-based methodologies, the project aimed to pave the way toward a cleaner and more sustainable energy future. The potential impact of such advancements on smart urban environments is immense.

The session was moderated by Ana Luísa Pereira, Innovation Manager at dstsolar, and kicked off with an opening by Sílvia Castro,  CMU Portugal Executive Director. Sílvia provided insight into the Program’s overall activities and the significance of the Large Scale Projects initiative, which has proven to be impactful in various ways. Building Hope was one of the 12 flagship projects selected under the 2018 competitive Call launched by COMPETE 2020, ANI, and FCT under the GO PORTUGAL initiative that was concluded in June 2023.

Mauro Costa, R&D Project Manager at dstsolar, took the stage and delivered a presentation, providing a comprehensive overview of Building Hope’s primary objectives, team achievements, and key responsibilities. The project was led by dstsolar in collaboration with  WATT-IS, InnovationPoint, S.A., Instituto Superior Técnico, Universidade de Coimbra, and the Electrical Engineering Department at Carnegie Mellon University.

Together, this consortium of expertise and knowledge has driven Building Hope toward its goal of revolutionizing energy management practices in smart urban environments. The HOPE platform has been extensively tested throughout the project in diverse contexts, including industrial, office, and retail buildings. These real-world trials have been instrumental in not only validating the platform’s capabilities but also showcasing its innovative features and the significant impact it can have on energy management.

To present the developed models, Henrique Pombeiro from the partner company Watt.IS focused on the outcomes of using data analytics with the presentation “Smarter and more efficient buildings through data analytics: contributions of load forecasting and non-intrusive disaggregation of consumption”.  The use of data analytics has proven to be a powerful tool in optimizing energy consumption and improving overall efficiency. Next, Carlos Santos Silva from Instituto Superior Técnico took the stage and delved into his team’s contributions, presenting “Flexibility, Performance, and SmartContracts: New services in intelligent building management.” His presentation focused on three critical aspects of the project: estimating hourly flexibility, predicting asset performance, and evaluating smart contracts. These focus areas are key to unlocking the full potential of smart urban environments and ensuring energy optimization. From Universidade de Coimbra, Ehsan Asadi delivered a compelling presentation entitled “Optimization of energy efficiency in buildings through the integration of flexibility resource integration.” This presentation showcased how integrating flexibility resources into building management can significantly enhance energy efficiency, making strides toward a greener and more sustainable future.

Finally, Ricardo Lopes, representing InnovationPoint, provided an in-depth exploration of the HOPE platform during his presentation. He highlighted the platform’s main features and showcased its overall performance. 

According to the project PI from dstsolar,  Ana Luísa Pereira “participating in this project was extremely important. It opened up the possibility for us to explore different areas with partners, valuing our investment in digital research projects associated with transition and energy efficiency. The cooperation between the teams of the different partners (dstsolar, Watt.IS, IST, UC and CMU) was enriching and allowed us to explore data processing and the interconnection of different areas of knowledge. We believe that these initiatives centered in the cooperation between entities of the academic and scientific system with companies, are essential for the development of new technologies and really innovative products.”

Giving a look into the project main outcomes, Ana Luísa Pereira refers that “the main advantage of the HOPE platform is that it is a tool to help reduce energy consumption in service, commercial or industrial buildings. During the project, a functional prototype of the platform was developed, but it will still be necessary to continue the development to reach a market-ready version. We are aware of the added value of this platform and its perfect alignment with dstsolar’s business strategy.” 

To conclude the session, the Building Hope consortia expressed their gratitude to SONAE, a prominent Portuguese multinational company that played a vital role in the project’s success. SONAE’s support involved providing their buildings for extensive testing, allowing the project to flourish and achieve its objectives. 

Catarina Nunes Soares, representing SONAE, took a moment to reflect on the impact of Building Hope on the company. She emphasized that the project had been a transformative experience, providing invaluable insights into the practices that can drive energy efficiency and sustainability within their facilities. The knowledge gained through this collaboration will not only enhance SONAE’s current buildings but will also inform the construction of future structures, ensuring a more eco-conscious approach.

As the session unfolded, it became evident that Building Hope’s journey has not only been one of innovation but also a testament to effective collaboration and the pursuit of positive change. The collective achievements of this project will undoubtedly serve as a beacon for future initiatives seeking to create a lasting impact on energy management and sustainability.

The Building Hope – Building Holistic Optimization of Prosumed Energy  – project is co-financed by the European Regional Development Fund (ERDF) through the Operational Program for Competitiveness and Internationalisation (COMPETE 2020) and by the Portuguese Foundation for Science and Technology (FCT) under the CMU Portugal Program.

On June 30th, the IntelligentCare project consortia gathered at Hospital da Luz Learning Health for the presentation of the CMU Portugal Large Scale Project results. The IntelligentCare project presents an innovative approach to managing multimorbidity (MM) by adopting a patient-centric perspective and striving for optimized allocation of hospital resources, thus promoting the transition to value-based healthcare.

The primary challenge involved developing a patient-centric solution by leveraging analytical methods and Artificial Intelligence (AI) to analyze data derived from electronic health records (EHR) and remotely reported measures such as patient-reported outcomes (PROMs), life events, quality of life, and physical activity. IntelligentCare, to capture additional value, incorporated intelligent sensors and mobile solutions to capture these variables, referred to as additional value variables (AVVs).

The ultimate goal was to create a comprehensive platform that aggregates all patient data, provides suggestions for medical interventions, and generates a risk score for hospitalization based on the patient’s condition.

Nuno André Silva, Head of the Intelligent Care and Complexity Science Center and the innovation area of ​​Hospital da Luz Learning Health delivered the opening remarks. In his introduction, he focused on the project’s objectives and highlighted the pressing challenges an aging population poses. One significant challenge emphasized was the rise in patients with multimorbidity (MM) and the resulting strain on the sustainability of the healthcare sector. Providing some context, he mentioned that according to the SNS (Portuguese Healthcare System), 38% of the Portuguese population is affected by this issue, leading to a greater demand for medical assistance and, consequently, increased healthcare utilization and costs.

The project consortia were represented at the meeting by members of all the involved Institutions, starting with Carlos Amaral of Priberam, who discussed data collection and NLP techniques. Mário Gaspar Silva, from INESC-ID, talked about disease phenotypes and the importance of early signaling and characterization of patients with MM. Plínio Moreno from Institute for Systems and Robotics (ISR-Lisboa) focused on smart sensing technology. Pedro Ferreira joined remotely from CMU (Heinz College of Information Systems and Public Policy) and discussed the extraction methodologies for clinical pathways developed along the project.  Jaime Machado, from Luz Saúde, presented the IntelligentCare platform that has been developed and which combines all the previous elements. And Bernardo Neves, Internal Medicine Doctor at Hospital da Luz Lisboa, showcased the clinical application with a real-world case study.

Pedro Ferreira, PI of the project at Carnegie Mellon, highlighted the benefits of this collaboration between CMU and Portuguese partners. “The IntelligentCare project brought together a unique multidisciplinary team, including faculty at CMU, faculty in Portuguese Universities, and practitioners in a novel and inspiring way. The leadership of Hospital da Luz was extraordinarily impactful and allowed us to focus the project on outcomes that matter for the hospital and patients. This project exemplifies how one can successfully publish state-of-the-art research immediately used in practice to improve people’s lives. I look forward to working with our Portuguese partners in subsequent research questions that our project identified”.    

To close the meeting, Francisca Leite, Director at Hospital da Luz Learning Health, discussed the future and its focus on the impact of AI in the Healthcare sector in the upcoming years. According to her, “The accomplishments of the IntelligentCare project will enable Hospital da Luz to revolutionize multimorbidity management. Integrating AI technologies will ensure every patient receives the care they need while reducing unnecessary costs and strains on our system. This project sets a pioneering example for the broader adoption of AI in healthcare, underscoring how technological innovation can positively impact patient care and system productivity. As we move forward, our goal is to continue incorporating AI across various healthcare domains, ensuring optimized resource allocation, early risk detection, and an overall more sustainable healthcare sector. We are thrilled to see what the future holds and are confident that AI will play a significant role in shaping a more sustainable, effective, and patient-centric healthcare sector in the coming years.”

The CMU Portugal large Scale Project IntelligentCare is led by Hospital da Luz Lisboa/Hospital da Luz Learning Health in collaboration with Priberam, INESC-ID, Institute for Systems and Robotics (ISR-Lisboa), Instituto Superior Técnico and the Heinz College at Carnegie Mellon University (CMU).

IntelliGentCare(LISBOA-01-0247-FEDER-045948) is co-financed by the European Regional Development Fund (ERDF) through the Operational Program for Competitiveness and Internationalization (COMPETE 2020) and by the Portuguese Foundation for Science and Technology (FCT) under the CMU Portugal Program.

 

The FLOYD project, a CMU Portugal Large Scale Collaborative initiative,  hosted a final Workshop on June 22nd at Altice Labs HQ in Aveiro and online. The event brought together members of the consortia and a diverse audience for a discussion on connected mobility and the integration of autonomous driving in our society.

The R&I project, led by Capgemini Engineering, aims to build a technological stack for offering high-performance network/computation services to vehicular users. The project has worked on improving mobility services and accelerating the integration of autonomous driving in society by advancing the networking and computation technological basis on which complex mobility and autonomy services and functionalities can be built. These include uninterrupted and seamless connectivity in diverse environments, leveraging edge and cloud computing, measuring network quality, control and prediction mechanisms, orchestration, security, and integration in various application use cases.

Funded through PT2020 and FCT under the CMU Portugal Program, the FLOYD – 5G/SDN Intelligent Systems For LOw latencY V2X communications in cross-Domain mobility applications initiative is led by Capgemini Engineering in collaboration with ALTICE LABS, S.A. , Instituto Superior de Engenharia do Porto,  Instituto de Telecomunicações, VORTEX Colab and the Computer Science Department at CMU. 

Through the collaboration between all these R&I industrial and academic entities, FLOYD has enabled the development, validation, and exploitation of various components and technologies, addressing multiple unanswered interrogations between vehicular and network communications.

“The collaboration between the different partners was extremely important, and each played its part. Capgemini was highly focused in the collection and analysis of 5G data, to make predictions to improve network performance and also with Edge Cloud concepts. Altice Labs made contributions to support the platoons between public and private 5G networks. Cister contributed with the use of Applications to assist the autonomous platoons and communications to 5G radio access points, IT for Wifi edge offloading and Multipath concept, Vortex in the security of the communications between the platoon members and CMU as deep app know-how.”, explains João Silva, Capgemini Engineering.

According to the project PI at CMU, Peter Steenkiste, Professor of Computer Science and of Electrical and Computer Engineering, “The FLOYD project was a great opportunity for both students and faculty in the CMU Portugal Program to work with industry research at Altice Labs and Capgemini Engineering.  The collaboration provided valuable insights into important challenges in using 5G technology for connected mobility. This led to interesting and relevant research projects in areas such as analyzing the latency of 5G communication, energy efficiency, and the overhead of the 5G control plane.”

Artificial Intelligence has been combined with breakthrough technologies in edge computing and 5G, which allowed the development of a framework of tools for networking and communications to support platooning applications. Platooning is an application where a group of vehicles moves one after another in proximity, acting jointly as a single physical system. The scope of platooning is to improve safety, reduce fuel consumption, and increase road use efficiency.

“It was a truly enriching experience to have the invaluable support and expertise of Carnegie Mellon University as well as the partners and entities involved in the challenging Floyd project. Through all the hard work and collective work, it was possible to make significant and promising advances in supporting autonomous driving in platoons, exploring topics such as 5G infrastructure, ‘Edge Cloud,’ or Artificial Intelligence. I hope that these important contributions and collaborations with CMU will serve as a basis for establishing promising partnerships in the future.”, adds João Silva, Capgemini Engineering.

In addition to disclosing the project’s outcomes, the workshop also featured presentations from other projects addressing related topics such as Route 25, 5G-Auto, InSecTT, and STEROID. 

Project FLOYD (POCI-01-0247-FEDER-045912) is co-financed by the European Regional Development Fund (ERDF) through the Operational Program for Competitiveness and Internationalization (COMPETE 2020) and by the Portuguese Foundation for Science and Technology (FCT) under the CMU Portugal Program.

You can learn more on the project’s website and through the video below.